Evaluation of Mayer-Rokitansky-Kuster-Hauser (MRKH) Patient Families by Whole Genome Sequencing

Abstract

Introduction: MRKH is a characterized by the congenital absence of the uterus and vagina in 46,XX individuals. A subset of these patients also has associated renal, skeletal, cardiac and/or auditory defects. Familial cases suggest a genetic component, but to date only pathogenic variants in WNT4 and HNF1B have been confirmed. We hypothesize that de novo heterozygous variants in candidate genes will be present in some patients with MRKH. Methods: DNAs from 30 quads (an MRKH proband and three relatives) were subjected to whole genome sequencing (WGS), and heterozygous variants in coding regions with < 0.02 frequency were filtered by two different methods. In the first approach, variants were filtered by 1) top consequence variant (splice site, stop-gain, frameshift, and missense, respectively); 2) impact score; 3) mapping quality; 4) cytobands; 5) intolerance; 6) de novo variants; and 7) plausibility based on familial genotype. The second approach considered only heterozygous variants found in the proband and absent in all other family members, which were then filtered by top consequence (splice donor and acceptor sites, stop-gain, frameshift). Results: Five pedigrees were excluded for inadequate sequence in one or more individuals. 55,033 variants in coding regions with < 2% frequency were identified in the 25 remaining quads for analysis. Using the first approach, 42 candidate gene variants in 32 genes were identified - 12 splice variants, 10 stop-gains, 15 frameshift variants and 5 missense variants. Of these, MUC22 contained 2 missense variants from different families. Additionally, DICER1 had multiple splice variants and is essential for mouse urogenital tract development. In the second approach, 39 candidate genes were identified—6 splice variants in 6 genes, 18 stop-gains in 17 genes, and 17 frameshift variants in 16 genes. Zinc finger genes (ZNF418, ZNF646, ZNF135, and ZNF772) comprised the most frequent class of the 39 genes. Two genes (MIR4436A and ZNF418) contained attractive variants in two different families. Conclusion: WGS has been shown to improve detection of gene variants in coding regions, more so than whole exome sequencing (WES). We previously performed WES on 111 MRKH probands without family members and analyzed variants in candidate genes suggested by mouse and preliminary human studies. Interestingly, in this study, only three genes overlapped with previously suspected candidate genes. Here, we identified new candidates based upon potential deleteriousness. These candidate genes will be studied further in our families to determine their role in Mullerian development.